A fast forward-looking super-resolution imaging method on a high-speed platform

Abstract Radar forward-looking imaging has raised many concerns in the civilian field. However, when employing forward-looking imaging methods on a high-speed platform, traditional Doppler compensation methods can result in significant compensation errors and exceed the Doppler compensation boundary. At the same time, the forward-looking geometric configuration results in the mutual coupling of range and Doppler information and the corresponding Doppler convolution matrix needs to be constructed per range cell. In addition, the large matrix dimensions of the convolution matrix result in enormous computational complexity. This paper proposes a fast super-resolution method based on high-speed platform Doppler compensation and a low-rank approximation batch processing (LRA-Batch) framework. First, the Doppler compensation matrix for high-speed platforms to eliminate Doppler centroid spatial variations in the range direction is constructed. Then, the Doppler convolution matrix only needs to be constructed once after compensation. The forward-looking imaging model under high-speed platforms has been improved through the above-mentioned method. Next, the sparse reconstruction optimization problem corresponding to the forward-looking super-resolution imaging model is reformulated into the matrix form, and the matrix dimension is reduced through low-rank approximation. Subsequently, the 2D echo data can be processed directly, avoiding the high computational complexity associated with line-by-line processing. The relevant simulation experiments to evaluate the performance of the proposed method have been designed. Simulation results prove that the proposed LRA-Batch method can obtain effective reconstruction results under low operational complexity.